JPH0366357A - Biosignal processor - Google Patents

Abstract

PURPOSE:To decrease identification errors by providing a neural computation means which reads out a memory means and identifies the waveform of a biosignal and a logical computing means. CONSTITUTION:The pulse wave signal of the blood circulation of the peripheral part can be obtd. if a finger tip 1 is irradiated with light from a photoirradiation section 3 and the quantity of the transmitted light is measured in a photodetecting part 4 by the light absorptivity of the hemoglobin in blood. This signal is passed through a filter section 5 and amplifier section 6 and is differentiated twice in a two-times differentiating section 7, by which the accelerated pulse wave signal is obtd. The accelerated pulse wave signal subjected to the differentiation twice is sampled by a sampling means 8 and the signal digitized by an AD converting means 9 is stored in a memory means 10. The logical computing means 11 reproduces the signal from the memory means 10 and inputs the reproduced accelerated pulse wave to the neural computation means 12 in which 7 patterns A to G are previously learned. Which pattern is the measured accelerated pulse wave is identified by this means. Since the waveform of the accelerated pulse wave signal is identified, the errors in the pattern identification are decreased.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of Industrial Application The present invention relates to a health device that detects biological signals from a living body or a part of a human body and identifies health conditions.

BACKGROUND OF THE INVENTION Conventionally, this type of biological signal processing device is as shown in Fig. 3(a), (
(b). FIG. 3(a) shows a signal inputted from the living body detection means 2 to the filter means 5, passed through the signal amplification means 6, inputted to the A, D conversion means 9 by the sampling means 8, and converted into a digital value. Storage means lO
, the biological signal waveform is reproduced by the logical operation means 11,
It is configured to measure waveform amplitude, inflection points, etc. and identify biological signals. FIG. 3(b) shows the same power as in FIG. 3(a), but the detected biological signal is manually input to the FFT analyzer 13, the power spectrum of the biological signal is measured, and the logical calculation means 11 is used to measure the power spectrum of the biological signal. This is a configuration for identifying signals.

Problems to be Solved by the Invention However, in the above configuration, the characteristics of a part of the waveform of the biological signal, that is, the amplitude of the inflection point, the frequency component, and the biological condition are identified. The problem was that the structure was not such that identification could be made using all the information included.

The present invention solves such conventional problems and aims to identify a biological condition with an amount of information as close as possible to the total amount of information contained in detected biological signals.

Means for Solving the Problems In order to solve the above problems, the biological signal processing device of the present invention includes means for detecting biological signals from biological systems, sampling means for sampling biological signals, and biological signals obtained by sampling. AD conversion means for converting from an analog value to a digital value; storage means for storing the digital signal obtained from the AD conversion means; and neural computation means for reading out the storage means and identifying and recognizing the waveform of the biological signal. It is equipped with the following.

Operation The present invention uses the above-mentioned FIFE to perform AD conversion on the detected biological signal, reproduces the waveform of the biological signal from the digitized biological signal, compares the waveform with a plurality of predetermined constant patterns, and determines which pattern Using neural computation means, we identify whether the waveform is close to the
Since the original waveform of the biological signal containing all amplitude and frequency information is pattern-identified, the biological condition can be checked with relatively few errors.

Embodiments Hereinafter, embodiments of the present invention will be described based on the accompanying drawings. In FIG. 1, l is a fingertip, 2 is a biological signal detection means, 3 is a light illumination part, 4 is a light receiving part, and the output of the biological signal detection means 2 is passed through a filter part 5 to an amplification part. 6
is input. The output of the amplifying section 6 is input to the twice differentiating circuit section 7, and the output thereof is sampled by the sampling means 8 and input to the AD[1 means 9. A storage means 10 stores the output of the AD conversion means 9. Reference numeral 11 denotes a logic operation means, which is coupled to the storage means 10 and the neural computing means 12.

In the above configuration, when the fingertip 1 is irradiated with light from the light irradiation section 3, the amount of transmitted light is measured by the light receiving section 4 due to the light absorption property of hemoglobin in the blood, as shown in FIG. 2(a). It is possible to obtain peripheral blood circulation pulse wave signals. By differentiating this signal twice through the filter section 5 and the amplifying section 6 in the twice differentiating circuit section 7, an accelerated pulse wave signal as shown in FIG. 2 (b) is obtained. It is generally known that this waveform can be divided into seven patterns depending on the biological condition. A in Figure 2 (C) is a waveform that is normally seen in energetic young people when blood circulation is good, and B in Figure 2 (C) is a waveform that is seen in the course of a process where blood circulation becomes insufficient. ,
Still in good condition. C in FIG. 2(C) shows a state where blood circulation has become insufficient. D, 已, F, G in Figure 2 (C)
indicates a fairly poor state of blood circulation. Generally, there are four inflection points 1-1, I-2, I-3, and I-4 in an accelerated pulse wave, and each of I-2, I-3, and I-4 is different from the height of I-1. Find the height ratio and use this ratio to
-G is known to be classified into seven patterns. However, in the present invention, the ratio is not calculated, but the twice differentiated accelerated pulse wave signal is sampled by the sampling means 8, and the signal digitized by the AD conversion means 9 is stored in the storage means 10. The logical operation means 11 is the storage means 1
0 and inputs the reproduced accelerated pulse wave signal to the neural computation means 12 which has previously learned the seven patterns A to G, and identifies which pattern the measured accelerated pulse wave signal corresponds to. With this configuration, pattern identification errors are significantly reduced because the waveform of the accelerated pulse wave signal itself is identified by neural matching, compared to the conventional method of determining the ratio and identifying which pattern is based only on the amplitude of the waveform. effective. Although the present embodiment has been described using an accelerated pulse wave signal as an example, it goes without saying that the present invention can also be applied to the identification of other measurement signals of biological systems.

Effects of the Invention As described above, the biological signal processing device of the present invention provides the following effects.

(1) Since the original waveform of the biometric signal itself is pattern-identified using pre-trained neural computation means, identification errors are significantly reduced compared to a configuration that identifies only amplitude or frequency components. be.

(2) Since the identification error is considerably reduced by the neural computation means, the number of patterns to be identified can be increased.

[Brief explanation of drawings]

FIG. 1 is a circuit block diagram of a biological signal processing device according to an embodiment of the present invention, FIG. 2 is a signal waveform diagram of each part of the device, and FIG.
FIG. 4 is a block diagram of a conventional biological signal processing circuit. 2...Biological signal detection means, 8...Sampling means, 9...AD conversion means, IO...
... Storage means, 11 ... Logical operation means, 1
2...Neural communication means.

Claims (1)

[Claims] a means for detecting a biological signal from a biological system; a sampling means for sampling the biological signal; an AD conversion means for converting the biological signal obtained from the sampling means from an analog value to a digital value; A biological signal processing device comprising a storage means for storing a digital signal obtained by the calculation, a neural computation means for reading out the storage means and identifying a waveform of the biological signal, and a logic operation means for controlling the neural computation means.